primary and bacterial production in sea ice in the

Vol. 20: 107-118, 1999 AQUATIC MICROBIAL ECOLOGY

Aquat Microb Ecol 1 Published December 15

Primary and bacterial production in sea ice in the northern Baltic Sea

Pia Haecky*, Agneta Andersson"

Department of Microbiology. UmeA University. 901 87 Umed. Sweden

ABSTRACT: The temporal variation of ice primary and bacterial production along with ice algal, bacterial and heterotrophic flagellate biomass were studied at a coastal station in the northern Baltic Sea throughout the ice-covered period of 1996 (January to April). Ice core samples were taken every week and analyzed for abundance and production of different microorganisms. In addition, physical and chemical parameters were measured. The ice algae were Limited by light during the first 3 mo of the study. The algal production showed a peak in the middle of April, which coincided with a marked increase in light availability. Shortly after that, the system became phosphorus depleted and primary production decreased rapidly. Bacterial biomass and production rates were relatively low and stable before the ice algal bloom. After the ice algal bloom, bacterial production increased rapidly, while the biomass remained low. The growth rate of small heterotrophic flagellates (<l0 pm), calculated from increase in biomass, was more than 1 order of magnitude higher than the bacterial production rate following the ice algal bloom. Thus, small heterotrophic flagellates were using food sources other than bacteria for growth after the ice algal bloom. On an annual basis, the ice algal and bacterial production accounted for < l % and <0.1% respectively of the total production (ice + pelagic) due to a short ice-covered season. During the ice-covered season, however, the ice algae accounted tor 10% of the total algal production, whlle ice bacterial production was 0.2 % of the total bacterial produckon.

KEY WORDS: Sea-ice . Primary production - Bacterial production - Heterotrophic flagellates . Nutrients . Light

INTRODUCTION

Sea ice microbial communities are known to occur in both polar and temperate ice-covered regions includ- ing the Baltic Sea area (Horner 1985, Palmisano & Gar- rison 1993, Norrman & Andersson 1994, Ikavalko &

Thomsen 1997, Haecky et al. 1998). The sea ice microbial comn~unity consists of representatives from all trophic levels of the microbial food web and includes bacteria, algae, heterotrophic protozoa and small met- azoa (Garrison 199 1). Sea ice microbes successfully inhabit the surface, interior and the bottom of the ice (Horner et al. 1992). Ice algal growth rates are controlled by brine temperature and salinity, as well as by

Present addresses: 'Marine Biological Laboratory (University of Copenhagen) Strandpromenaden 5, 3000 Helsing~r. Denmark. E-mail: [email protected]

"Marine Ecology, Department of Ecology and Environrnen- tal Science, Umed University, 901 87 Umeb, Sweden

the availability of light and nutrients (Cota et al. 1991). During ice formation hypersaline brine is produced, in which nutrients and other dissolved constituents of the seawater are concentrated. Subsequently, the brine drains into milli- to micrometer-sized pockets and channels throughout the ice matrix, where the chemical composition is controlled by sea ice temperature (Weeks & Ackley 1986). Brine volume decreases with temperature, and salts contained within ~t are concentrated correspondingly. Bulk nutrient concentrations in newly formed sea ice are essentially equal to those in the water at the time of ice formation, while nutrient concentrations in older sea ice are influenced by brine drainage and by the activity of sea ice microbial communities (Dieckmann et al. 1991, Gleitz et al. 1995). The brine, which is the environment actually experienced by sea ice microorganisms, is thus funda- mentally different from the pelagic environment. Brine channels contain large surface areas per brine volume. Salinity and nutrient concentrations can be extremely

Q Inter-Research 1999

108 Aquat Microb Ecol20: 107-1 18, 1999

high but variable, depending on age and temperature of the ice. On the other hand, the light environment in sea ice habitats is more stable than that in planktonic habitats, since ice is not subject to large vertical displacements in the irradiance field. As a result, the sea ice provides an environment where ice algae can grow, while phytoplankton growth 1s severely limited by light, thus prolonging the growth season by at least 1 to 3 mo (Cota et al. 1991). Dunng the winter months irradiances are low, which leads to light limitation and low light adaptation of ice algae. However, in late spring and early summer, light may cause photoinhibi- tion of ice algae (Kirst & Wiencke 1995). In annual sea ice, algae are limited by light at the beginning of the season and by inorganic nutrients later in the season, as these are usually consumed or drained out of the ice (Gosselin et al. 1990, Cota et al. 1991).

The length of the ice-covered season, as well as the growth season of sea ice algae, increases with latitude. At low latitudes the ice-covered season is short (1 to 3 mo) and it usually occurs during the darkest period of the year, while polar regions can be covered by ice all the year round. The contribution of the production within the ice, relative to the total (ice + pelagic) production is thus expected to vary through the season and with latitude. In recent studies in the Arctic and Antarc- tic Oceans, ice algae accounted for 3 to 57 % and 20 to 33% of total annual primary production, respectively (Legendre et al. 1992, Kirst & Wiencke 1995, Gosselin et al. 1997). In the central Arctic Ocean ice algal production during the summer months was found to con- tribute up to 57% of the entire primary production, while the algal contribution was only 2 to 5 % in the sur- rounding regions. In an early study in the Beaufort Sea, sea ice production was estimated to constitute as much as 2/3 of primary production during winter and early spring (Horner & Schrader 1982). Thus, ice productivity, relative to pelagic productivity, shows a marked variation within and between different sea areas.

Every year, the northern Baltic Sea is ice-covered for 4 to 6 mo, whi1.e the southern Baltic Sea is more or less ice-free all yea.r around. Diverse microbial communities have previ.ous1.y been found in the interior brine channels of the annual land-fast ice of the northern Baltic Sea (Norrman & Andersson 1994, Ikavalko 1997), but the bacterial and algal production have not been measured before. In this study, we measured the magnitude of sea ice bacterial and algal production in order to estimate its contribution to the total production in the low productive reglon of the northern Baltic Sea. Sea ice productivity was compared to pelagic productivity at the same location during the ice-covered season and during the whole year. A carbon budget model of microbial carbon flows was created from measured stocks and rates.

MATERIAL AND METHODS

Sampling. Ice samples were collected weekly at a coastal station in the Gulf of Bothnia (63' 33' 76" N, 19' 50' 74" E) throughout the ice-covered season, January to April 1996 (Fig. 1). On 2 occasions (13 and 26 March 1996), 6 stations were sampled along a 10 km west- southeast transect from the coastal high frequency station to a station further offshore (63" 29' 30" N, 19" 56' E) . A stainless steel corer, 13 cm diameter (Franson, LuleA Technical High School, Sweden), powered by a motor, was used for ice sampling. Triplicate ice cores were routinely taken, measured and cut into 3 equal length sections (top, middle and bottom of the ice core). However, during the transect samplings ice cores were kept intact (12 March) or cut into 2 ~ q i l a l length sections (top and bottom of the ice core, 26 March). The ice fractions were transported back to the laboratory in acid-washed buckets, while keeping them dark and cold. In the laboratory, the samples were left for 10 to 20 min in order to collect drainage. The drainage consisted mainly of brine, since the time of collection was too short for any notable melting of the ice. This fraction will henceforth be called brine. The brine was kept at O°C until further processing. Brine from the triplicate samples was pooled to a final volume of at least 80 m1 from each section. This was the minimum volume needed for the measurements of algal and bacterial production, chlorophyll a (chl a) , bacterial and heterotrophic flagellate numbers and

I

l o o t o o 30'

Fig. 1. Map of the Baltic Sea. 0: sampling site

Haecky & Andersson Primary and bacterial production in sea ice 109

inorganic nutrients. After collecting the brine, the re- mainder of the ice core was completely thawed for the determination of chl a, bacteria and heterotrophic flagellates. The remaining ice core was processed as soon as it was fully thawed and, like the brine, always kept at 0°C. Whole ice core concentrations were determined by adding concentrations in the brine and the remaining ice cores in correct volumetric proportions, which are henceforth called ice concentrations. Stock and production values are given per ice volume or in area1 units Data on pelagic variables were retrieved from a coastal monitoring station situated close to the high frequency station. At this station, samples from 0 to 20 m depth were taken throughout the year every 2 to 4 wk.

Nutrients and salinity. Nutrient samples were taken from the brine of the ice core sections and filtered through precombusted (5 h at 400°C) Whatman GF/F filters. The samples were analyzed immediately or frozen until analysis. Nutrient concentrations were determined using a TRAACS auto analyzer (Alfa Lava1 Bran and Luebbe TRAACS 800) and standard seawater procedures (Grasshoff et al. 1983). The nutrient concentrations are reported per volume of brine. Mea- suring inaccuracies (standard deviations of triplicates) for phosphate, nitrate + nitrite, ammonium, and silicate concentrations were 2.6, 1.2, 6.9 and 1.3%, respectively. Salinity was measured in brine water and in residual thawed ice cores, using a salinometer (AGE Instruments Inc., Model 2100).

Light. Photosynthetically active radiation (PAR, 400 to 800 nm) was measured hourly in air and at each sampling occasion under the ice, using a spherical quantum meter (Li-Cor, Nebraska, USA). PAR under the ice was measured through a hole drilled in the ice (15 cm), with the sensor attached to a curved rod. The rod kept the sensor in an upright position under the ice at a distance of -40 cm from the drilled hole.

Chlorophyll a and algae. For chl a determinations, 100 rnl samples were filtered gently (5100 mm Hg) onto 25 mm Whatman GF/F filters. The chl a on the filters was extracted in the dark at room temperature in 95% ethanol for 24 h without grinding. Fluore- scence was measured in a Perkin Elmer fluorometer (LS 30) and chl a was calculated according to HEL- COM (1988). The average coefficient of variation based on triplicates for chl a measurements was 12 %. Samples for algal species analysis were fixed with Lugol's solution at a final concentration of 2 %, concentrated in a sedimentation chamber and studied in an inverted microscope at 200 to 400x magnification.

Primary production. Primary production was measured by the 14C technique (Parsons et al. 1984). Tripli- cate light and dark samples (5 ml), consisting of brine from the top, middle and bottom of the ice cores, were

poured into polycarbonate tubes (Nalgene). Brine samples, rather than whole ice, were used for production measurements since sea ice microbial communities live within the brine. An inocculum of 0.64 pCi of carner- free sodium (I4C) bicarbonate (0.1 mCi mmol-') was added to the samples, and they were incubated for 3 to 4 h around noon in the ice at the sampling station. To create in situ light and temperature conditions, samples were incubated in ice holes drilled to sampling depths (centre of the top, middle and bottom sections), and covered with ice and snow. After incubation, HCl was added to a final concentration of 0.4 M to remove ex- cess 14C-bicarbonate. The samples were bubbled with air for 30 min and left overnight to free the samples of residual '4C-bicarbonate. Scintillation cocktail (Opti- phase 'Highsafe' 111) was added, and the samples were counted in a scintillation counter (Beckman LS1801). The daily assimilation values were calculated by multiplying the measured primary production rates with the ratio between total daily insolation and insolation during the incubation period. The primary production in the ice core was calculated by multiplying the production in the brine with a chl a factor [(chl a ice)/(chl a brine)], since during drainage only a fraction of the brine and organisms contained within it was recovered. The average coefficient of variation based on triplicates for primary production measurements was 11 %. Pel- agic primary production measurements were performed as described in Andersson et al. (1994).

Heterotrophic bacterial and flagellate biomass. Formaldehyde was added to ice brine samples, samples from wholly melted ice cores, and water column samples (final concentration 1.5 %). For the determination of bacterial biomass a 2 to 15 m1 sample was filtered onto black 0.2 pm polycarbonate filters, stained with acridine orange, mounted in paraffin oil, and examined by epifluorescence rnicroscopy (Hobbie et al. 1977). Estimates of bacterial cell volumes were acquired by image analysis (Blackburn et al. 1998).

Samples for counting heterotrophic flagellates from the ice (5 to 20 ml), were stained with acridine orange at a final concentration of 2.4 ppm and filtered onto black 0.8 pm polycarbonate filters. By using this weak staining, the heterotrophic microorganisms could be distinguished from the autotrophic ones by the ab- sence of autofluorescence (Andersen & Sarensen 1986). At least 50, but in most cases 100, small flagellate cells ( < l 0 pm) per slide were counted at 1250x magnification. Between 10 and 100 medium-sized flagellate cells (210 pm) per slide were counted at 250x magnification. The sizes of all cells were measured under the microscope using a calibrated ocular micrometer. The cell volumes were calculated assuming simple geometric shapes approximating those of the organisms. Small- and medium-sized heterotrophic

110 Aquat Microb Ecol20: 107-118, 1999

flagellates from the water column were fixed with acid February was the coldest month, with temperatures Lugol's solution at a final concentration of 2%, and 10 frequently dropping below -15°C. After 12 April, air to 50 m1 of the sample was concentrated in a sedimen- temperatures were generally >O0C. The water temper- tation chamber. One chamber diameter was scanned ature under the ice (0 m) was -0.11 to -0.23OC from in an inverted microscope at 400x magnification using January through March, which is close to the freez- phase contrast, and heterotrophic flagellates were ing temperature of brackish water with 4%0 salinity identified by morphology and coloration. The cell car- (-0.24"C; Fig. 2b). After 12 April the water tempera- bon content of bacteria and flagellates was calculated from cell volume: pgC cell-' = 0.125 X cell volume (pm3) (Pelegri 0.8

a) p Ice thickness et al. 1999). 0.7 --

Bacterial production. Bacterial pro- 0.6 --

duction was measured by the tritia- - 0.5 --

ted 3H-thymidine incorporation method g 0.4 -'

(Fuhrman & Azam 1982), modified by 0.3 --

centrifl~gation of the samples. Triplicate 0.2 -r

brine samples and controls were incu- 0.1

bated in darkness at in situ tempera- 0

tures. The saturation level of 3H-thymi- 4 dine (specific activity, 82 Ci mmol-') -

0 ------ uptake was measured at the beginning g

-4 of the sampling season (19 January 2

1996). 3H-thyrnidine was added to brine -8 samples in the range of 5.8 to 43 nM, and 2 g -12 the saturation level was found to be 23.3 nM. This concentration was added -16

to bacterial production samples over the -20 - whole sampling period, assuming a con- $ 600

stant saturation level. The pelagic sam- $ 500 d

ples were treated as described above, 3 400 but with the addition of 25 nM 3H-thyrni- - 300 dine. Bacterial production was calcu- 2 lated using the empirical conversion fac- 200

L:

tor 1.5 X 1018 cells mol-' 3~-thymidine 3 100

incorporated, which has been estimated 0 18

for the northern Baltic Sea (Wikner & - 30 16 Hagstrom in press). The bacterial pro- ' V)

14 duction in the ice core was calculated by E 25

12 multiplying the production in the brine 20

m with a bacterial biomass factor [(ice bac- 10 , - 2 15 terial biomass)/(brine bacterial biomass)]. 8 3. The average coefficient of variation based y 10 6 C

on triplicates for measured bacterial pro- M 4 2 5

duction rates was 14 %. 2

0 0 s z 5 z g v

. ! a k J 2 2 & $ " & 2 Z E Z E E Z E 4 4 9 4 Q

RESULTS " E ~ ~ ~ " ~ ~ m ~ ~ ~ w m '0

The 'hdied area was covered with 35 Fig. 2. (a) Thickness of the ice and the snow depth at the sampling site dunng to 70 cm of sea ice from January to April the ice covered season 1996. (b) Temperature in the air and the water under and from the end of January to the mid- the ice (0 m), and (c) the average incident irradiance during the daylight

die of ~ ~ r i l there was 6 to 13 cm snow on hours. Daylight hours were defined as the hours where the irradiance was >7 pm01 quanta m-' S-' (d) Light under the ice (I,) was calculated according the ice (Fig' 2a)' The average air to Beers law: I, = I. e-" where I. is the intensity at the surface and k is the

temperature between -20 and extinction coefficient. I. was measured hourly during the sampling period; +4"C during the study period (Fig. 2b). k was calculated from measured light intensities under the ice

Haecky & Andersson: Primary and bacterial product~on in sea ice 111

ture increased to >0.5"C. The average lncident irradiance during daylight hours increased from -20 pm01 quanta m-2 S-' in the beginning of January to -600 pm01 quanta m-' S-' at the end of April (Fig. 2c). The number of daylight hours increased steadily through the sampling period from 4 to 17 h (Fig. 2d). The average light intensity under the ice varied between 2 and 18 pm01 quanta m-2 S-' in January. Due to increasing ice thickness and the development of a snow cover, light under the ice dropped to 0.2 - 6 pm01 quanta m-2 S-' in February to March. In April the snow disappeared and the average light under the ice increased up to 30 pm01 quanta m-2 S-' for 16 h d-l.

The water column salinity was relatively stable during the sampling period (-4 %o) , while the brine and the ice salinity varied between 2-12 and 0.2-1.6%0, respectively (Fig. 3). The ice and the brine salinity generally decreased during the study, due to gradual desalination. Desalination of sea ice is commonly caused by gravity drainage and flushing by surface melt-water (Maykut 1985). Between 30 January to 20 February and 2 to 9 April brine salinity increased, indi- cating that the brine had been freezing. The inorganic

Fig. 3. Salinity of the (a) brine, (b) ice and (c) the water under the ice at 2 m depth

nutrient concentrations in the brine were normalized to seawater salinity (4.3%0), to correct for dilution or concentration during melting or freezing. Normalized concentrations of phosphate, nitrate + nitrite and ammonium in the brine were generally higher than those in the water column (Fig. 4a,b,c). However on 16 April, the nutrient concentrations in brine from the bottom and middle of the ice reached a minimum and were slmilar to those in the water column. Silicate concentrations were similar in the brine and in the water column during the whole sampling period (Fig. 4d).

Fig. 4. Concentrations of (a) dissolved inorganic phosphate (DIP), (b) nitrite + nitrate, (c) ammonium and (d) silicate in brine from the top, middle and bottom of the ice and in the water under the ice (2 m). The brine nutrient concentrations

are normalized to seawater salinity (4.3%)

112 Aquat Microb Ecol20: 107 -118, 1999

Production rates of the ice were cal- Table 1. Comparison of the relative proportion of collected and total brine in the

,-dated from the production in the brine sea ice at the high frequency coastal station. The algal (chl a) and bacterial blo-

multiplied by ice biomasshrine biomass concentrations in collected brine are given in % of ice (brine + residual ~ c e ) biomass concentrabons

mass, assuming that algal and bacterial specific activities were the same in the ice remaining after brine removal. If part of the microbial biomass was frozen into the ice or selectively retained within the ice during drainage due to large cell size, the collected microorganisms would not be representative of the whole ice microbial community. The total brine volume in the ice was calcu-

Collected Brine Collected brine Bacterial Chl a in brine volumea biomass in collected

collected brine brine (% of ice volume) (% of brine volume) (% of ice biomass)

Average 1.3 11.2 11.0 10.4 4.7 Mm. 0.2 5.2 3.2 2.2 0.1 Max. 4.9 20.7 52.2 36.1 14.7

'Calculated from brine salinity and temperature (Maykut 1985)

lated from salinity and temperature of ice (Maykut 1985), and compared to the amount of collected brine (Table 1). The proportion of the bacteria! biomass drained out of the ice was on average 10% of the total bacterial biomass, which was close to the amount of brine drained out of the ice (1 1 %). This indicates that bacteria were not frozen into the ice matrix,

Ice

and were readily drained out of the ice. However, only -5 O/o of the ice algal biomass was collectec! during b-me drainage, as opposed to the 11 % of brine collected. This indicates that half of the ice algae were either frozen into the ice matrix, or 'stuck' in the ice pores during drainage (large cells that were unable to pass through thin brine

channels). The method we used to calculate whole ice primary and bacterial production

2 rates might thus have overestimated ice pri- a) Ice mary production rates, while bacterial pro-

-- OTop duction rates were closer to the true values.

Middle Both chl a concentration in the ice and ice algal primary production per area were rela-

D 2 tively stable and low from January to February, % $ 0.8 -- while both increased from 20 February to

6 March (Fig. 5a,b). In March, chl a concentration and primary production were again stable at a somewhat higher level. Concurrent with

b) increasing light intensities and day lengths, an

0.5 -- ice algal bloom occurred in the beginning of c 0 . .- - April, as indicated by peaks in the chl a con- 3 $ 0.4

-- U N

centrations and primary production. After the g E 0.3 -- ice algal bloom, chl a concentration and pri- P 2 mary production decreased rapidly. The 2 g 0.2 -- .- & 5 phytoplankton spring bloom was observed af-

0.1 -- -e ter the ice break-up, but had already started during April under the ice (Fig. 5c).

0.0 -d 3 Water under the ice The ice algal community was mainly com-

c) posed of diatoms, but the autotrophic dino- - 40 -- - + -Chla $ 7 flagellate Pendiniella catenata (Levander)

g -- + b a y pductio" Balech and the mixotrophic microflagel-

3 - late Dinobryon faculiferum (Wille'n) Wille'n z 6 20 -- ? S (Chrysophyceae) were relatively common.

10 -- 0,5 X 5 Small unidentified autotrophic flagellates were also frequently observed. Navicula pel-

C C C D D " ~ ~ L h ~ ~ b b agica Cleve was found throughout the sam- z z ~ 9 b ~ z z ~ < < < < < pling season and dominated during the ice al- \ D s W a s a n - 88flrnz~rn C\I N gal bloom. Nitzschia frigida Grunow., Melosira

arctica (Ehrenberg) Dickie, Chaetoceros wig-

Fig. 5. (a) Chlorophyll a, and (b) primary production rates in the sea ice, hamii Brightwell and Navicula vanhOeffenii and (c) in the water column, integrated from 0 to 20 m depth Gran were also commonly found.

Haecky & Andersson: Pnmary and bacterial production in sea ice 113

and production rates varied by a factor of 6 and 65, respectively, while areal rates varied by a

The bacterial abundance varied Table 2. Average values, data range (within parentheses) and coeffic.ients of

from -7 104 to 2.7 105 cells rnl-1 in vanation (CV) of ice algal and bacterial biomass and production from 2 transect

the ice, 4.2 105 to 2,3 106 cells ml- l in samplings on 13 and 26 March 1996. The transects consisted of 6 stations, from the nearshore, high frequency station to a station situated 10 km offshore

the brine and between 7.5 X 105 and

factor of 3 and 6, respectively. This indicates

1.3 X 106 cells ml-' in the water column (data not shown). The ice bacterial biomass remained relatively stable during the whole sampling period (Fig. 6a). The bacterial production in the ice was relatively high at the beginning of Jan- uary, while it was low during February and March (Fig. 6b). In April the bacterial production rate increased rapidly, following the ice algal bloom. The

Fig. 6. (a,b) Bacterial biomass and production (calculated from 3H-thy- that the temporal variations of bacterial bio-

midine uptake rates) in the ice. (c) Bacterial biomass and production in mass and production were higher than spatial

13 March 1996 26 March 1996 Average C V ('%) Average CV (%)

- -

Chl a (mg m-') 1.3 (0.7-1 9) 31 1.3 (0 6-2.2) 44

primay production 200 (63-360) 57 85 (49-120) 35 (pm01 Cm-2 d-')

Bacterial biomass 81 (46-152) 46 (pm01 C m-')

Bacterial production 1.3 (0.5-2.9) 73 (l-1mo1 C ln-' d-')

- the water column, integrated from 0 to 20 m depth variations in the studied area.

pelagic bacterial biomass was stable throughout the sampling season, while the pelagic bacterial production was slowly increasing system, bacterial production per biomass varied be- (Fig. 6c). The bacterial production per biomass varied tween 0.01 and 0.12 d-' during the whole sampling pe- between 0.006 and 0.05 d-' from January to March, and riod. Thus, the specific growth rates of ice bacteria were after that it increased rapidly up to -0.25 d-' (16 to 22 higher than bacterial growth rates in the water column April) and finally to 0.38 d-' (29 April). In the pelagic during and after the ice algal bloom.

Spatial variations of primary and bacterial production, chl a and bacterial biomass in ice were determined during transect samplings (13 and 26 March). The transects consisted of 6 stations, from the nearshore, high frequency sampling station to a station situated 10 km offshore (Table 2). The coefficients of variation (CV) of the spatial variation for chl a and primary production were high (31 to 44 and 35 to 57 %, respectively) relative to the coefficients of variation of triplicate samples within ice cores (12 and 11 % respectively). However, the temporal variations of chl a and primary production varied by a factor of 6 and 65 respectively between minimum and maximum values, while areal chl a concentrations and primary production rates varied by a factor of 3 to 4 and 2.4 to 6 respectively. This indicates that the temporal variations of chl a and primary

1 production were higher than spatial variations in the studied area. The transect sampling on

0.8 .g -- 26 March indicated high spatial variation of e 2

D - - - - 0 - - 0.6 2 wu bacterial production rates by the high CV for - a- - Biomass E production rates measured during the transect

0.4 2 (73 %) relative to the average CV of triplicate g g samples within ice cores (14 %). However, tem-

0.2 S poral variations between minimum and maxi-

o mum values of the measured bacterial biomass

114 Aquat Microb Ecol 20: 107-118, 1999

The biomasses of small- and medium-sized Table 3. Monthly Ice algal and bactenal production durlng the Ice cov-

heterotrophic flagellates were relatively low ered season 1996 (data range wlthin parentheses). The contribution of

until the beginning of April (Fig. On 1ce production to the total production (pelagic + ice) in the studled sea area is also shown

April, during the peak in primary production, the biomass of both small- and medium-sized flagellates started to increase. Both size classes of flagellates reached their peaks on 16 April after the primary production maximum. The heterotrophic flagellates in the water column under the ice increased simultane- ously with the flagellates in the ice (Fig. ?c). The abundance of small- and medium-sized heterotrophic flagellates in the ice were between 8.5 X 10' to 5.9 X 103 and 3 to 16 cells m l l respectively (data not shown). The heterotrophic flagellate component was rlnminated by unidenMied small heterotrophic flagellates. Medium-sized heterotrophic flagellates were

0.60 a) Ice

<l0 pm n 0.40 -- OTop

Middle M Bottom

0.20 -- - E l U - g 0.00

Ice prlmary Ice bacterial Bacterial production production production in O/o of primary

(mm01 C ' m - ) (mm01 C m-') production p - -

January 0.51 0.06 11.9 (0.7-24 8) February 1.35 0.01 0.7 (0.5-3.0) March 2.40 0 02 0.7 (0 5-0.9) April 5.37 0.33 6.2 (0 7-106) Sum 9.54 0.42 4.4

Ice production compared to total production (ice + water column)

( " ,> ) ("/.l Ice-covered season 10.1 0.18 Annual 0.38 0.06

Fig. 7. Biomass of (a) small ( < l 0 pm) heterotrophic flagellates and ( h ) medium-sized (210 pm) heterotrophic flagellates in the ice, and (c) in the water under the ice integrated from 0 to

20 m

composed of dinoflagellates, and unidentified flagellates. The biomass concentrations of algae and heterotrophic flagellates were generally higher in ice than in the water column, whereas the biomass of bacteria was lower in the ice than in the water column (data not shown). Calculated on an area1 basis, the bacterial, algal and flagellate biomasses were considerably lower in the ice than in the water column (Figs. 5, 6 & 7) .

In order to estimate the significance of ice production in the studied area, the production values in the ice and in the water column were compared during the ice-covered season and over the whole year (Table 3). The ice primary production accounted for 10% of the total production during the ice-covered season, while bacterial production in the ice accounted for 0.2% of the total bacterial production over the same period. On an annual basis, ice primary and bactenal production were < l and <0.1% respectively of the total production.

DISCUSSION

The maximum ice algal biomass was approximately the same as that observed in a study performed at the same location during 1994 (1.5 mg chl a m-', Haecky et al. 1998), but lower than that found during winter 1990 (10 mg chl a m-2, Norrman & Andersson 1994). The main difference between the study in 1990, with high ice algal biomass, and the 2 other studies was the nutrient concentration in the water column. In 1990, phosphate in the water column was 1.6 pM at the beginning of the ice-covered season, but in 1994 and 1996 the concentration was only 0.12 and 0.18 pM respectively (Fig. 4; Andersson et al. 1994, Haecky et al. 1998). The potential ice algal biomass in 70 cm sea ice from the given phosphate concentration would be 40, 3.0 and 4.5 mg chl a m-* in 1990, 1994 and 1996, respectively

Haecky & Andersson: Primary and bacterial production in sea ice

(calculated by using the Redfield molar C:P ratio of 106, and the conversion factor 3 m01 C g-' chl a; Red- field et al. 1963, Haecky et al. 1998). The actual ice algal biomass concentrations were less than half of the potential concentrations, where the difference was most likely caused by brine drainage and predation. From the 3 studies in northern Baltic Sea ice , it can be concluded that the ice algal biomass accumulation depends on the amount of nutrients (particularly P) trapped in the ice during the ice algal bloom period. The ice algal biomass in the northern Baltic Sea was low compared to those found in, for example, Arctic and Antarctic sea ice (Legendre et al. 1992, Kirst & Wiencke 1995). Also, nutrient concentrations measured in the water column in this study were low compared to winter nutrient concentrations in the Atlantic and Pacific Oceans (Sakshaug 1989), resulting in a lower potential ice microbial biomass.

Light, rather than nutrients, limited ice algal growth during the first 3 mo of the investigation, since the phosphate concentration in the brine, before the ice algal bloom, was higher than k, measured for mixed natural phytoplankton populations (average

k, = 0.27 PM, range = 0.05 to 0.5 PM; Cembella et al. 1984). The marked increase of primary production in response to increasing light availability, in the beginning of April, also showed that ice algae were light limited (Figs. 2d & 5b). Decreasing phosphate concentrations between 9 and 16 April might have been the cause of the crash of the ice algal bloom. Mortality of polar diatoms can be induced by nutrient exhaustion in combination with high pH, oxygen oversaturation, and low CO2 concentration, which are conditions that are regularly attained during advanced stages of ice algal blooms (Giinther et al. 1999).

Good correlation between bacterial and primary production was found from the end of January to 9 April (r2 = 0.91). Similar trends have been observed in sea ice from the western Baltic Sea and Antarctica, where a strong linkage between primary and bacterial production was observed (Kottmeier et al. 1987, Mock et al. 1997). During the spring bloom, bacterial production constituted 3 to 5 % of the ice algal primary production in Resolute Passage, High Arctic (Smith & Clement 1990). However, bacterial production exceeded primary production in sea ice studied during autumn and

26 March 2 April 9 April 16 April 22 April

Flagellates (2 10 W)

Algae

W C dL detritus

Bacteria

Flagellales (< 10 W)

Fig. 8. Carbon budget of microbial processes in sea ice during and after the ice algal bloom (26 March and 22 April, 1996). Car- bon stocks of algae, bacteria and heterotrophic flagellates are shown in boxes and production rates are shown by arrows. Algal carbon was calculated from chl a (3 m01 C g-' chl a), and carbon flow rates from production rates and changes in carbon stocks. The growth efficiency of bacteria and heterotrophic flagellates was assumed to be 40 %. Medium-sized heterotrophic flagellates (210 pm) were assumed to graze on algae, and small heterotrophic flagellates (< l0 pm) on bacteria. In cases where the carbon demand of small heterotrophic flagellates exceeded the bacterial carbon production, ingestion of DOC and detritus is suggested

116 Aquat Microb Ecol20: 107-1 18, 1999

winter (Kottmeier et al. 1987, Rivkin et al. 1989). In the algae were the major source of dissolved organic car- present study, total bacterial production during the ice- bon (DOC), and concentrations up to 480 mm01 C 1-' covered season equaled 4 % of the ice primary produc- were found (Smith et al. 1997). Metazoan grazing was tion, which was similar to the spring situation described most likely responsible for part of the ice algal bio- above (Table 3). During April part of the bacterial pro- mass lost from the ice after the ice algal bloom, since duction was either grazed or lost from the ice, since the these have been found in relatively high concentra- bacterial production increased rapidly, while bacterial tions in the lowermost section of sea ice from the biomass remained relatively stable (Fig. 6a,b). A similar northern Baltic Sea (Synchaeta baltica, Norrman &

situation was observed in ice from Forbisher Bay Andersson 1994). (Canadian Subarctic), where the increase of bacterial Bacterial production alone was insufficient to sustain standing stock was low compared to bacterial cell pro- the observed growth of small heterotrophic flagellates duction (Bunch & Harland 1990). just after the ice algal bloom, since the carbon demand

A budget of the known carbon pools and flows dur- was more than 1 order of magnitude higher than the bac- ing and after the ice algal bloom is outlined in Fig. 8. terial production rate (Fig. 8). Moreover, the estimated Following the ice algal bloom, the algal biomass carbon demands of flagellates are minimum estimates, decreased, while algal growth increased rapidly. The since they a r e based on net changes in numbers and resulting loss of ice algal biomass could only partly be losses of flagellates were not accounted for in the present accounted for by rapidly increasing heterotrophic fla- budget. This indicates that the heterotrophic flagellates gellate biomasses (Fig. 7a,b). The carbon demand of used carbon sources other than bacteria for growth (e.g. medium-sized heterotrophic flagellates between 2 to grazing on small primary producers, viruses or direct up- 29 April accounted for 2 to 11 % of the algal loss rate, take of DOM). In the Canadian Arctic, the bacterial graz- calculated from biomass accumulation and assuming ing rate seemed sufficient to meet heterotrophic flagel- a growth efficiency of 40% (Table 4; Sanders et al. late carbon demand before the ice algal bloom, but 1992). The carbon demand of medium-sized heter- became inadequate after the ice algal bloom (Laurion et otrophic flagellates was similar to findings from the al. 1995). The authors suggested that the consumption of Canadian Arctic, where microprotozoan carbon con- small algae and direct ingestion of DOM became more sumption was estimated to be 1 to 8% of the net bio- important after the ice algal bloom. Thus, the het- mass loss from the ice algal population (Sime-Ngando erotrophic flagellates in sea ice seem to be able to utilize et al. 1997). The specific growth rates of the micropro- more diverse food sources than flagellates from the tozoan populations in sea ice from the Canadian Arc- pelagic system. tic were also similar to the growth rates estimated in The contribution of ice algal production to the total the present study (Table 4). The remaining loss of primary production during the ice-covered season was algal biomass, not consumed by flagellates or bacte- 10% (Table 3). In Arctic regions, where the ice cover ria, was considered to accumulate in a detrital pool, was 55 to 90 % of the surface area, the ice algal contn- which potentially might have resulted in high concen- butions to total primary production rates were close to trations of dissolved organic matter (DOM). In bottom those found in the present study (2 to 5%, Gosselin et ice from the Canadian Arctic and northern Japan, ice al. 1997). A significantly higher contribution of the ice

algal production to the total production

Table 4 . Growth rate and net production rate of heterotrophic flagellates, cal- in the ice-c0vered

culated from biomass changes during and after the ice algal bloom central Arctic Ocean, where the ice algal contribution was 57 %. In the present study, the contribution of the total ice algal production to the annual water column production was 0.4%, which was considerably lower than values reported from the Arctic first year ice (3 to 25%, Legendre et al. 1992). The main reasons for higher ice algal production in annual sea ice from the Arctic Ocean are the longer ice- covered season compared to that of the northern Baltic Sea, and higher nutn- ent con.centration in the water during ice formation, resulting in higher potential ice algal biomass (see above)

pd Net production rateb (cl-') (pm01 C m-2 d-l)

2-9 Apr 9-16 Apr 9 Apr 16 Apr

Small TOP 0.08 0.41 1.9 169.7 heterotrotrophic Middle 0.28 0.07 7.9 3.4 flagellates Bottom 0.09 0.17 2.8 17.1 ( < l 0 pm) Sum 12.5 190.3 Medium-sized Top 0.10 0.06 6.0 5.0 heterotrotrophic Middle 0.04 0.12 1.6 11.0 flagellates Bottom 0.10 0.11 1.5 3.7 (210 P) Sum 9.0 19.8

'Growth rate p (d-') = (h N,- In No) / t . N, was flagellate biomass at time t , No at time 0 b ~ e t production rate (pmol C m-2 d-') = p (d-') X flagellate-C (pm.01 C m-')

-

Haecky & Andersson: Primary and bacterial production in sea ice 117

Highly productive bottom ice algal communities have not been observed in the northern Baltic Sea, but are assumed to occur under much of the Arctic Sea ice, thus increasing the ice algal production. However, maximal ice algal biomasses in the northern Baltic Sea have been observed to vary over 1 order of magnitude (see above). In addition to that, the annual pelagic primary production rates vary by a factor of 6 among different years in this sea area (1984 to 1996: 1 to 5.6 m01 C m-2 yr-' (Nordstrom & Wikner 1996). This suggests that the contribution of the ice algal production to the total annual production might vary at least 1 order of magnitude.

The contribution of the ice bacterial production to the total production was << 1 % both on an annual basis and during the ice-covered season. In Antarctic Sea ice (McMurdo Sound), the ice bacterial production accounted for 3 to 5 % of the total bacterial production in platelet ice and for 1 to 2 % in congelation ice during late winter and early spring. The reason for the low ice bacterial production in this study was presumably grazing of bacteria as indicated by the relatively abun- dant presence of small heterotrophic flagellates.

It seems that the microbial community in sea ice from the northern Baltic Sea developed through 3 phases: from January to March a light-limited, low productive winter community developed. During the first half of April a spring bloom community with rapidly increasing primary productivity developed, followed by a post bloom comnlunity, where heterotrophic processes dominated. Ice bacteria seemed to play only a minor role in the carbon turnover, while heterotrophic flagellates, particularly after the ice algal bloom, played a major role. However, to acquire a deeper understand- ing of the ecological processes within the sea ice of the Gulf of Bothnia, more work is needed on the production and turnover of the organic carbon pool as well as the nutrition of heterotrophic flagellates.

Acknowledgements. This work was supported by grants from the Swedish National Sciences Research Council (B-AA/BU 06554-304), and from EU (BASYS MAS3-CT96-0058). We thank Birgitta Karlsson, Knstina Samuelsson and Jonas Wester for help with the field work, and Erik Lundberg and Carl-Henrik Stangenberg for chemical analysis. Furthermore we would like to thank Sheila and Henry Blackburn for lin- guistic corrections, and Tom Fenchel for help with the micro- scopic work. The laboratory facilities and the use of data from the pelagic monitoring program at Umea Marine Sciences Center are gratefully acknowledged.

LITERATURE CITED

Andersen P, Ssrensen H (1986) Population dynamics and trophic coupling in pelagic microorganisms in eutrophic coastal waters. Mar Ecol Prog Ser 33:99-109

Andersson A, Haecky P, Hagstrom A (1994) Effect of temperature and light on the growth of micro-, nano- and pico-

plankton: impact on algal succession. Mar Biol120: 511-520 Blackburn N, Hagstrom A, Wikner J , Hansson RC, Bjarnsen

PK (1998) Rapid determination of bacterial abundance, biovolume, morphology, and growth by neural network- based image analysis. Appl Environ Microbiol 64: 3246-3255

Bunch JN, Harland RC (1990) Bacterial production in the bottom surface of sea ice In the Canadian subarctic. Can J Fish Aquat Sci 4?:1986-1995

Cembella A, Antia N, Harnson P (1984) The utilization of inorganic and organic phosphorus compounds as nutrients by eucaryotic microalgae: a mulitdisciplinary per- spective. CRC Crit Rev Microbiol 10:31?-391

Cota GF, Legendre L, Gosselin M, Ingram RG (1991) Ecology of bottom ice algae: I . Environmental controls and vari- ab~lity. J Mar Syst 2.257-277

Dieckmann GS, Lange MA, Ackley SF, Jennings JC Jr (1991) The nutrient status in sea ice of the Weddell Sea during winter: effects of sea ice texture and algae. Polar Biol 11: 449-456

Fuhrman J , Azam F (1982) Thyrnidine incorporation as a measure of heterotrophic bacterioplankton production in marine surface waters: evaluation and field results. Mar Biol 66:109-120

Garrison DL (1991) Antarctic sea ice biota. Am Zoo1 31:17-33 Gleitz G, Rutgers van der Loeff M, Thomas DN, Dieckman

GS, Millero FJ (1995) Comparison of summer and winter inorganic carbon, oxygen and nutrient concentrations in the Arctic sea ice brine. Mar Chem 51:81-91

Gosselin M, Legendre L, Therriault JC, Demers S (1990) Light and nutrient limitation of sea-ice microalgae (Hudson Bay, Canadian Arctic). J Phycol26:220-232

Gosselin M. Levasseur M, Wheeler P, Horner R, Boot B (1997) New measurements of phytoplankton and ice algal production in the Arctic Ocean. Deep-Sea Res 44:1623-1644

Grasshoff K, Ehrhardt M, Kremling K (1983) Methods of seawater analysis, 2nd revised and extended version. Verlag Chemie, Weinheim

Giinther S, Gleitz M, Dieckman GS (1999) Biogeochemistry of Antarctic sea ice: a case study on platelet ice layers at Drescher Inlet, Weddell Sea. Mar Ecol Prog Ser 177:l-13

Haecky P, Jonsson S, Andersson A (1998) Influence of sea ice on the composition of the spring phytoplankton bloom in the northern Baltic Sea Polar Biol 20:l-8

HELCOM (Helsinki Commission) (1988) Baltic Sea environment proceedings no. 27 D. Guidelines for the Baltic monitoring programm for the third stage. Baltic Marine Envi- ronment Protection Commission, Helsinki Commission, Hels~nki

Hobbie JE, Daley RJ, Jasper S (1977) Use of nucleopore filters for counting bacteria by fluorescence rnicroscopy. Appl Environ Microbiol33:1225-1228

Horner RA (1985) Sea ice biota. CRC Press, Boca Raton, FL Horner R, Schrader GC (1982) Relative contribution of ice

algae, phytoplankton, and benthic rnicroalgae to primary production in nearshore regions of the Beaufort Sea. Arc- tic 35.485-503

Horner RA, Ackley SF, Dieckman GS, Gulhksen B, Hoshiai T, Legendere L, Melnikov IA, Reeburg WS, Spindler M, Sul- livan CW (1992) Ecology of sea ice biota. 1. Habitat, temi- nology, and methodology. Polar Biol 12:417-427

Ikavalko J (1997) Further observations on flagellates within sea ice in northern Bothnian Bay, the Baltic Sea. Polar Biol 19:323-329

Ikavalko J , Thomsen HA (1997) The Baltic Sea ice biota (March 1994): a study of the protistan comnlunity. Eur J Protisto133:229-243

Aquat Microb Ecol20: 107-118, 1999

Kirst GO, Wiencke C (1995) Ecophysiology of polar algae. J Phyco131:181-199

Kottmeier ST, Grossi SM, Sullivan CW (1987) Sea ice microbial communities. VIII. Bacterial production in annual sea ice of McMurdo Sound, Antarctica. Mar Ecol Prog Ser 35: 175-186

Laurion I, Demers S, Vezlna AF (1995) Th.e microbial food web associated with ice the algal assemblage: biomass and bacterivory of nanoflagellate protozoans in Resolute Passage (High Canadian Arctic). Mar Ecol Prog Ser 120: 77-87

Legendre L, Ackley SF, Dieckman GS, Gulliksen B, Horner R, Hoshiai T, Melnikov IA. Reeburgh WS, Spindler M, S u b - van CW (1992) Ecology of sea ice biota. 2. Global significance. Polar Biol 12:429-444

Maykut G (1985) The ice environment. In: Horner RA (ed) Sea ice biota, CRC Press, Boca Raton, FL, p 21-82

Mock T, Meiners KM, Giesenhagen HC (1997) Bacteria in sea ice and underlying brackish water at 54' 26' 50" N (Baltic Sea, Kiel Bight). Mar Ecol Prog Ser 158:23-40

Nordstrom A, Wikner J (1996) The biology of the water column. In: Forsgren G, Wiklund K (eds) The Gulf of Bothnia 1996-annual report from the marine monitoring pro- gramme (in Swedish). Umed Marine Sciences Centre. Lule.3 Alltryck, Lulea, p 11-12

Norrman B, Andersson A (1994) Development of ice biota in a temperate sea area (Gulf of Bothnia). Polar Biol 14: 531-537

Palmisano AC, Garrison DL (1993) Microorganisms in Antarc- tic sea ice. In: Friedman IE (ed) Antarctic microbiology. Wiley-Liss, New York, p 167-218

Parsons TR, Maita Y. Lalli CM (1984) A manual of chemical and biological methods for seawater analysis. Pergamon Press, Toronto

Editorial responsibility: Tom Fenchel, Helsingor, Denmark

Pelegri SP, Dolan J , Rassoulzadegan F (1999) Use of high temperature catalytic oxidation (HTCO) to measure carbon content of microorganisms. Aquat Mlcrob Ecol 16: 273-280

Redfield AC, Ketchum BH, Richards FA (1963) The influence of organisms on the composition of sea water. In: Hill MN (ed) The sea. Wiley, New York, p 26-77

Rivkin RB, Putt M, Alexander SP, Meritt D, Gaudet L (1989) Biomass and production in polar planktonic and sea ice microbial communities: a comparative study. Mar Biol 101: 273-283

Sakshaug E (1989) The physiological ecology of polar phytoplankton. In: Rey L, Alexander V (ed) 6th Conf Com Arct Int, 13-15 May 1985. EJ Brill, Leiden, p 61-89

Sanders RW, Caron DA, Berninger UG (1992) Relationships between bacteria and heterotrophic nanoplankton in marine and fresh waters: an inter-ecosystem comparison. Mar Ecol Prog Ser 86:l-14

Sirne-Ngando T, Juniper SK, Demers S ( l W 7 ) Ice-brine and planktonic microheterotrophs from Saroma-ko Lagoon, Hokkaido (Japan): quantitative importance and trophody- namics. J Mar Syst 11:149-161

Smith REH, Clement P (1990) Heterotrophlc activity and bacterial productivity in assemblages of microbes from sea ice in the high Arctic. Polar Biol 10:351-357

Smith REH, Gosselin M, Kudoh S, Robineau B, Taguchi S (1997) DOC and its relationship to the algae in the bottom ice communities. J Mar Syst 11:71-80

Weeks WF, Ackley SF (1986) The growth, structure, and properties of sea ice. In: Understeiner N (ed) Geophysics of sea ice. Plenum Press, NATO AS1 Series 3, p 9-164

Wikner J, Hagstrom A (in press) Bacterioplankton intra-annual variability: importance of hydrography and competition. Aquat Microb Ecol

Submitted: July 13, 1999; Accepted: October 28, 1999 Proofs received from author(s): December 7, 1999

primary and bacterial production in sea ice in the

Documents